Özalp Özer @Stanford University, MS&E 262 April 29, 2015 Forecast Sharing Game
2
Announcements
Please sit as sparsely as possible.
Check your email to ensure that you received the link.
Keep your completed assignment – turn it in after the game is over.
3
The Procurement Game
1. Retailer observes the actual value of X and sends a report to the Supplier
2. Supplier receives the Retailer’s report and determines production quantity Q
3. Demand = X + Y is realized
4. Supplier delivers MIN(Q, Demand) to the Retailer, which are then sold to the end customers
8
Let’s Play the Game
Try your best to finish making your decision in each round within 120 seconds.
Keeping to the time limit will ensure we finish the game on time.
Remember to click “Continue” after you finish making decision/reviewing results in a round.
A decision of “0” means that you have been matched with a dummy player.
9
Debriefing of the Experiment
How did you report as a retailer?
How did you determine production quantity as a supplier?
What affected your decisions in the experiment?
11
Results – UTD Students R
epo
rt
Red line: 45 Degree line
Actual X Report P
rod
uce
Q
Red line: Optimal Q based on the retailer’s report
Trustworthiness Trust
Source: Özer and Zheng. 2011. Trust in Forecast Information Sharing. Management Science
12
Results – Stanford UG Students R
epo
rt
Red line: 45 Degree line
Actual X Report P
rod
uce
Q
Red line: Optimal Q based on the retailer’s report
Trustworthiness Trust
Source: Özer and Zheng. 2011. Trust in Forecast Information Sharing. Management Science 57(6), pp. 1111–1137.
13
Results – Columbia Students R
epo
rt
Red line: 45 Degree line
Actual X Report P
rod
uce
Q
Trustworthiness Trust
Source: Özer 2014. Forecast Information Sharing Game at Columbia GSB.
14
Results – Compared to Stanford Students
Report vs. Actual Value of X
Stanford Students You? (UTD)
Red line: 45 Degree line
15
Results – Compared to Stanford Students
Production Quantity vs. Report
Stanford Students You (UTD)
Partial trust and cooperation between the two parties!
Red line: Optimal production quantity based on the retailer’s report
16
What IF?
How would you change your strategy
If cost were lower?
If market uncertainty were higher?
If you are playing with your close friend/family?
?
High versus Low Cost of Overage
Lower cost induces trust and cooperation
Actual X
Rep
ort
Report P
rod
uce
Q
18
US - China trade accounts for 3% of the world trade volume
China at the cross roads of global supply chains
Several anecdotal evidence (e.g., Ebay vs Taobao)
Two ends of the cultural and institutional spectrum
U.S. China
Global Supply Chains
19
Cross Country and Within Country Results Forecast Inflation (Trustworthiness) Production Adjustment (Trust)
Trust and trustworthiness are lower in China (twice)
Chinese (and US) show higher trust towards U.S. partners
US-US supply chain > Cross-Country > CN-CN (10% less)
US Supplier
CN Supplier
0
10
20
30
40
50
US RetailerCN Retailer
21
42
27
48
US Retailer
CN Retailer
0
10
20
30
40
50
60
US SupplierCN Supplier
21
36
36
56
Source: Özer and Zheng. 2014. Trust, Trustworthiness, and Information Sharing in Supply Chains Bridging
China and the U.S. Management Science, 6(10) pp 2435-2460.
21
F(Q*) = Area of shaded region = 1/5
Supplier’s Decision– Newsvendor! (BUT?)
Demand uniformly distributed between 25 and 475 (?)
p = $100, c = $80
Cu = $20, Co=80
Critical Ratio = (20)/(100) = 1/5
475
f(D)
Demand
1/450
25 Q* = 115
Mean = 250
22
What is Trust?
What do you think “trust” means …
in the context of forecast sharing in a supply chain?
in a more general context?
What are the prerequisites for you to trust?
23
Trust in Forecast Sharing
Supplier’s trust:
Willingness to rely on the retailers report to determine production quantity
Retailer’s trustworthiness:
Disutility of deception controls incentive for forecast distortion
24
Definition and Determinants of Trust
“Trust is a psychological state comprising the intention to accept vulnerability based upon positive expectations of the intentions or behavior of another.”
Uncertainty and vulnerability are essential in defining trust
Determinants:
Risk attitudes
Aversion to betrayal
Reputation
Others? Quote from: Rousseau, D., S. Sitkin, R. Burt, C. Camerer. 1998. Not so different after all: A cross-discipline view of trust. Academy of Management Review, 23(3) 393-404.
25
How to effectively share information?
IT is necessary but not sufficient Firms may have incentive to manipulate forecasts/information to gain
advantage (i.e., induce the supplier to build more capacity, stock more inventory, …)
Contracts to align incentives based on pecuniary payoffs.
“Put your money where your mouth is”
Trust and Trustworthiness (non-pecuniary)
Shared Forecasts(Information): Field Data
26
1999 Q1 1999 Q2 1999 Q3 1999 Q4 2000 Q1 2000 Q2 2000 Q3 2000 Q4 2001 Q1 2001 Q2 2001 Q3 2001 Q4
Year and Quarter
1999 Q1
1999 Q2
1999 Q3
1999 Q4
2000 Q1
2000 Q2
2000 Q3
2000 Q4
2001 Q1
Actual
Source: M. Cohen, et al. 2003. Measuring Imputed Costs. Management Science, 49(12) 1653–1670.
27
Implications for Forecast Management
Soruce: Özer, Zheng and Chen. 2011. Trust in Forecast Information Sharing. Management Science 57(6), pp. 1111-1137.
28
Summary
Forecasting is a fundamental Prerequisite for production planning, material procurement,
capacity planning
Firms in a supply chain may have different information Retailers know consumer tastes better
Producers know the features of new products better
An efficient supply chain should take advantage of the better information
A key issue is how to ensure credible info sharing when conflicting incentives exist Align pecuniary (contracts) and non-pecuniary (trust) values